"Garbage In, Garbage Out." But Who's Catching It?

Emily Goh · 2026-04-14

Everyone knows the rule. Nobody's applying it.

"Garbage in, garbage out."

You've heard it a thousand times. You probably nodded along when someone said it in a meeting. It's technically correct. It sounds like something smart people know.

But here's the problem: most teams have no idea what's actually going in.

I've talked to dozens of ops teams. Finance teams. RevOps teams. People responsible for the data that runs their business. And when I ask them "How do you catch bad data?" -- the answers are almost always the same. They're also almost always terrifying.

Let me walk through the nine ways ops teams actually find bad data. And why every single one of them is too late.

1. Someone notices it in a meeting.

The CEO is presenting to the board. Pipeline forecast shows $3.2M in Stage 4. A board member asks "What's the win rate on those deals?" The VP of Sales looks at the number and realizes he has no idea. He pulls the deals up. Three of them are actually closed and already in the GL. One was deleted six months ago but the opportunity record is still there. Another one has a contact person who left the company three weeks ago.

This is how you catch bad data: by accident, in real-time, in front of people.

Cost: Credibility. A nervous board member. A follow-up call asking for a corrected forecast.

2. A customer gets mad.

Your customer success team sends out an NPS survey. The survey goes to the wrong person, or it gets bounced because the email address is old. The customer -- the good customer -- thinks you've gone inactive. They're annoyed. They don't renew.

You catch it during a churn postmortem when someone asks "How did we lose that account?" and the CSM realizes the contact record was three years out of date.

Cost: Customer relationship. Potential churn. Hours of CSM time investigating.

3. Month-end reconciliation surfaces gaps.

Your finance team is closing the books. They're reconciling revenue from Salesforce against the GL. The numbers don't match. There's $87K of revenue in the GL that isn't in Salesforce. $43K of deals marked "closed-won" in Salesforce that never made it to the GL.

Someone spends four hours in a spreadsheet figuring out which records are which.

Cost: Four hours of finance team time. A delayed close. Risk that the auditors notice something.

4. A failed integration breaks silently.

Your Zapier automation between HubSpot and Salesforce stopped syncing three days ago. Nobody noticed. You're missing the first 48 hours of follow-up on fresh leads.

You find it when someone does a random check and realizes "Hey, why haven't we gotten any leads in three days?" It actually broke two weeks ago. You're down 30 leads already.

Cost: Lost leads. Sales team wondering why pipeline dried up. Time troubleshooting the integration.

5. Someone manually audits a slice of data for compliance.

Your compliance person is pulling together documentation for a SOX audit. They spot-check 50 closed-won deals in your CRM. Three of them have wrong close dates. Two are missing the signed contract. One has conflicting customer names between Salesforce and the invoice.

Cost: Audit risk. Potential adjustment. If it's systematic, you might have to do a full reconciliation.

6. A new team member asks a dumb question.

Your new hire comes in and asks "How do I know if this customer record is the single source of truth?" And you realize you don't actually have a good answer. You dig in. You find four different contact records for the same company. The opportunity list has 17 old deals that were never closed-lost.

Cost: Time onboarding. Confusion. Loss of trust in the data.

7. A campaign fails and nobody knows why.

You segment your customer base and send a marketing campaign. The open rate is 8%. Your email partner emails you: "Heads up, 15% of your emails bounced." The other half went to bad emails that never got validated.

Cost: Wasted marketing spend. Low campaign performance. Someone has to figure out which segment was actually valid.

8. A manual data check finds something weird.

Your ops person is getting ready for a QBR. They spot-check 20 deals. Five are in the wrong stage. Three have no contact person assigned. Two are duplicated.

They spend two hours cleaning up what they can. But they never know how bad the full dataset actually is.

Cost: Hours of manual work. Limited confidence in the data.

9. An executive makes a decision on bad data and it blows up.

A director looks at customer churn data and decides to cut support costs because "churn is only 2%." But the churn metric is based on records that haven't been updated in six months. Your actual churn is 8%. A quarter later, you realize you've understaffed support.

Cost: Missed business targets. Damaged customer relationships. Hiring and onboarding costs.

What all nine have in common.

Every single one of these is reactive.

You find the bad data only after it's caused a problem. After it's affected a decision. After it's cost you time or money or credibility.

The fix.

You need to know what's actually in your data before you use it for decisions.

Not manually. Not through random spot checks. Systematically. Continuously.

A process that checks: Are these contacts real? Are these opportunities actually in the right stage? Are there duplicates? Are there orphaned records? Are there data inconsistencies between systems?

Not after something breaks. Before.

The tools you're using -- Salesforce, HubSpot, NetSuite -- they're not built for this. They're built to store data and let you access it. They're not built to validate that the data is any good.

That's the gap.

Gohvio is free to try. Upload your data and see what's hiding in it → gohvio.io